Nuala Brady

Local Contrast in Natural Images: Normalisation and Coding Efficiency

The visual system employs a gain control mechanism in the cortical coding of contrast whereby the response of each cell is normalised by the integrated activity of neighbouring cells. While restricted in space, the normalisation pool is broadly tuned for spatial frequency and orientation, so that a cells response is adapted by stimuli which fall outside its ‘classical’ receptive field. Various functions have been attributed to divisive gain control: in this paper we consider whether this output nonlinearity serves to increase the information carrying capacity of the neural code. 46 natural scenes were analysed with the use of oriented, frequency-tuned filters whose bandwidths were chosen to match those of mammalian striate cortical cells. The images were logarithmically transformed so that the filters responded to a luminance ratio or contrast. In the first study, the response of each filter was calibrated relative to its response to a grating stimulus, and local image contrast was expressed in terms of the familiar Michelson metric. We found that the distribution of contrasts in natural images is highly kurtotic, peaking at low values and having a long exponential tail. There is considerable variability in local contrast, both within and between images. In the second study we compared the distribution of response activity before and after implementing contrast normalisation, and noted two major changes. Response variability, both within and between scenes, is reduced by normalisation, and the entropy of the response distribution is increased after normalisation, indicating a more efficient transfer of information.

My left brain and me: a dissociation in the perception of self and others.

We investigated hemispheric asymmetries in face processing using a task in which participants judged the likeness of chimeric faces to their own face and to the face of a close friend based on their memory for those faces. When asked to choose which of two mirror-symmetric images (one made from the left half and one from the right half of a photograph of their face) looked more like themselves as remembered, participants showed a significant bias for the composite corresponding to the half face that lies in their right visual field when they look at themselves in the mirror. They showed the opposite bias when asked to make the same choice for images of a close friend, that is, they showed a significant bias for the composite corresponding to the half face that lies in their left visual field when they look at their friend. This result shows that in the case of these highly familiar faces--self and friend--the perceptual asymmetry is preserved in the memory representation. Assuming that people remember their own face as a mirror-image, the data also suggest a dissociation in face processing such that the left brain is dominant for the recognition of self and the right brain is dominant for the recognition of others.

Self-face recognition is characterized by “bilateral gain” and by faster, more accurate performance which persists when faces are inverted

We examine interhemispheric cooperation in the recognition of personally known faces whose long-term familiarity ensures frequent co-activation of face-sensitive areas in the right and left brain. Images of self, friend, and stranger faces were presented for 150 ms in upright and inverted orientations both unilaterally, in the right or left visual field, and bilaterally. Consistent with previous research, we find a bilateral advantage for familiar but not for unfamiliar faces, and we demonstrate that this gain occurs for inverted as well as upright faces. We show that friend faces are recognized more quickly than unfamiliar faces in upright but not in inverted orientations, suggesting that configural processing underlies this particular advantage. Novel to this study is the finding that people are faster and more accurate at recognizing their own face over both stranger and friend faces and that these advantages occur for both upright and inverted faces. These findings are consistent with evidence for a bilateral representation of self-faces.

Differing roles for the dominant and non-dominant hands in the hand laterality task

Determining the handedness of visually presented stimuli is thought to involve two separate stages—a rapid, implicit recognition of laterality followed by a confirmatory mental rotation of the matching hand. In two studies, we explore the role of the dominant and non-dominant hands in this process. In Experiment 1, participants judged stimulus laterality with either their left or right hand held behind their back or with both hands resting in the lap. The variation in reactions times across these conditions reveals that both hands play a role in hand laterality judgments, with the hand which is not involved in the mental rotation stage causing some interference, slowing down mental rotations and making them more accurate. While this interference occurs for both lateralities in right-handed people, it occurs for the dominant hand only in left-handers. This is likely due to left-handers’ greater reliance on the initial, visual recognition stage than on the later, mental rotation stage, particularly when judging hands from the non-dominant laterality. Participants’ own judgments of whether the stimuli were ‘self’ and ‘other’ hands in Experiment 2 suggest a difference in strategy for hands seen from an egocentric and allocentric perspective, with a combined visuo-sensorimotor strategy for the former and a visual only strategy for the latter. This result is discussed with reference to recent brain imaging research showing that the extrastriate body area distinguishes between bodies and body parts in egocentric and allocentric perspective.

Compensation is unnecessary for the perception of faces in slanted pictures

In four experiments, we explored the perception of facial distortions seen in pictures viewed from the side or from above or below. In all four, however, we disguised the slant of the picture surface by using a double-projection technique that removed binocular and monocular cues: Faces were digitized, distorted to mimic a particular slant behind the image plane, cropped to a frame, and presented to viewers for their judgments. In the first experiment, we found that simulated rotations around a horizontal axis (pictures seen as if from above or below) created more noticeable distortions in faces than did simulated rotations around a vertical axis (pictures seen as if from the left or right). In the second experiment, pursuing a result from the first but with a between-subjects design, we found that pictured faces with a slant around a vertical axis of 22° were seen as having no more distortion than unslanted faces. In the third experiment, we placed each image within a frame slanted either in the same way as or differently from the picture, and found no effect of frame. In the fourth experiment, we determined that viewers had little ability to match appropriately slanted frames with slanted pictures. Thus, we claim that part of the reason why one can look at moderately slanted pictures without perceptual interference is that the distortions in the image are subthreshold, or perhaps within the bounds of acceptability. These results contrast with the generally accepted theory that viewers mentally compensate for distortions in moderately slanted pictures.

Perception of ‘Best Likeness’ to Highly Familiar Faces of Self and Friend

We investigated the idea that our memory for familiar faces involves an accurate representation of their unique spatial configuration and, further, whether this configuration may be caricatured in memory. In separate experimental blocks, thirty-five Irish participants were presented with a series of photographic images of their own face and of the face of a close friend, and were asked to choose the image which looked most like themselves or their friend. Both sets of images included an original full-face colour photograph, and photographic distortions ranging from a highly caricatured (+100%) to a highly anti-caricatured (–100%) version of the original, generated with reference to newly created average male and female Irish faces. Contrary to suggestions that we hold a slightly caricatured version of a familiar face in memory, the mean ‘best-likeness’ image, calculated across both self and friend trials, was an anti-caricature of −13.88% which was significantly different from 0 (t69 = −5.34, p < 0.0001). The difference in the mean ‘best-likeness’ image chosen for self (–12.06%) and friend (–15.7%) was not significant (t34 = 0.715, p = 0.48). These results are discussed with reference to our ability to discriminate facial shape, together with the possibility that we idealise the attractiveness of faces of those close to us.

Rapid Perceptual Switching of a Reversible Biological Figure

Certain visual stimuli can give rise to contradictory perceptions. In this paper we examine the temporal dynamics of perceptual reversals experienced with biological motion, comparing these dynamics to those observed with other ambiguous structure from motion (SFM) stimuli. In our first experiment, naïve observers monitored perceptual alternations with an ambiguous rotating walker, a figure that randomly alternates between walking in clockwise (CW) and counter-clockwise (CCW) directions. While the number of reported reversals varied between observers, the observed dynamics (distribution of dominance durations, CW/CCW proportions) were comparable to those experienced with an ambiguous kinetic depth cylinder. In a second experiment, we compared reversal profiles with rotating and standard point-light walkers (i.e. non-rotating). Over multiple test repetitions, three out of four observers experienced consistently shorter mean percept durations with the rotating walker, suggesting that the added rotational component may speed up reversal rates with biomotion. For both stimuli, the drift in alternation rate across trial and across repetition was minimal. In our final experiment, we investigated whether reversals with the rotating walker and a non-biological object with similar global dimensions (rotating cuboid) occur at random phases of the rotation cycle. We found evidence that some observers experience peaks in the distribution of response locations that are relatively stable across sessions. Using control data, we discuss the role of eye movements in the development of these reversal patterns, and the related role of exogenous stimulus characteristics. In summary, we have demonstrated that the temporal dynamics of reversal with biological motion are similar to other forms of ambiguous SFM. We conclude that perceptual switching with biological motion is a robust bistable phenomenon.

Spatial Scale Interactions and Image Statistics

In natural scenes and other broadband images, spatial variations in luminance occur at a range of scales or frequencies. It is generally agreed that the visual image is initially represented by the activity of separate frequency-tuned channels, and this notion is supported by physiological evidence for a stage of multi-resolution filtering in early visual processing. The question whether these channels can be accessed as independent sources of information in the normal course of events is a more contentious one. In the psychophysical study of both motion and spatial vision, there are examples of tasks in which fine-scale structure dominates perception or performance and obscures information at coarser scales. It is argued here that one important factor determining the relative salience of information from different spatial scales in broadband images is the distribution of response activity across spatial channels. The special case of natural scenes that have characteristic ‘scale-invariant’ power spectra in which image contrast is roughly constant in equal octave frequency bands is considered. A review is presented of evidence which suggests that the sensitivity of frequency-tuned filters in the visual system is matched to this image statistic, so that, on average, different channels respond with equal activity to natural scenes. Under these conditions, the visual system does appear to have independent access to information at different spatial scales and spatial scale interactions are not apparent.

Is human motion detection subserved by a single or multiple channel mechanism

Two recent versions of a single channel model of motion perception have had impressive success in explaining direction discrimination by human observers for spatially filtered noise images in two-flash apparent motion. It has been argued that the dramatic breakdown in motion perception which occurs when one image in the two-flash sequence is low-pass filtered can be explained only by a single channel model. We show that neither version of the single channel model which has been proposed can explain performance for noise images chosen to provide comparable stimulation in the spatial channels known to subserve human vision. A multi-channel model of motion perception has little difficulty in explaining these results.

Shared or separate mechanisms for self-face and other-face processing? Evidence from adaptation.

Evidence that self-face recognition is dissociable from general face recognition has important implications both for models of social cognition and for our understanding of face recognition. In two studies, we examine how adaptation affects the perception of personally familiar faces, and we use a visual adaptation paradigm to investigate whether the neural mechanisms underlying the recognition of one’s own and other faces are shared or separate. In Study 1 we show that the representation of personally familiar faces is rapidly updated by visual experience with unfamiliar faces, so that the perception of one’s own face and a friend’s face is altered by a brief period of adaptation to distorted unfamiliar faces. In Study 2, participants adapted to images of their own and a friend’s face distorted in opposite directions; the contingent aftereffects we observe are indicative of separate neural populations, but we suggest that these reflect coding of facial identity rather than of the categories “self” and “other.”